Conventional search systems may include logic to detect and group related documents returned in response to a query. This logic tends to be fixed and pre-defined. This logic also tends to rely solely on text-based comparisons (e.g., string, proximity). Thus, these conventional systems may compare document content, the URL of documents, and/or subject lines of emails to determine whether documents are to be treated as part of a group.